The Hispanic US population increased by 61% in the last decade of the 20th century and, by2050, Hispanics may comprise 25%of the population. Nearly 80% of Hispanics aged e 20 years in the US are overweight or obese compared to 66% of non-Hispanic adults. This is a concern because obesity is a major risk factor for many diseases and also decreases lifespan. However, our knowledge of the obesity-mortality association, whether derived from body mass index (BMI: kg/m2) or direct measures of body composition, is based primarily on data from persons of European ancestry. Thus, little is known about the obesity-mortality association among Hispanics. To fill this gap in understanding, we propose two specific aims: (1) To derive estimates of the association of BMI to both hazard ratios (HR) and years of life lost (YLL) among Hispanic adults by analyzing raw data from at least nine large datasets, each containing over 500 Hispanic respondents, and each of sufficiently long follow-up, and (2) To estimate HRs and YLLs, as a function of obesity determined by measures of body fat and waist and hip measurements contained in the datasets. We will analyze each of the datasets separately, as well as by pooling the data. Conducting the data analysis in both ways allows us to observe the sensitivity of the estimates to sampling variability due, in part, to study-specific and sample-related factors and provide overall estimates that incorporate all data. [unreadable] [unreadable] Project Narrative: Given the increasing number of Hispanics in the US population coupled with the high prevalence of overweight and obesity among this minority group, a greater understanding of the obesity-mortality association among Hispanics is vital to the public health. Such information will help in making funding and policy decisions on research and on obesity treatment and prevention programs among Hispanics. The aims of this study are also consistent with NIH's mission to address health disparities and to improve the health status of minority groups. [unreadable] [unreadable] [unreadable]